x86_64: Don't use softirq safe locks in smp_call_function
authorAndi Kleen <ak@suse.de>
Sat, 21 Jul 2007 15:10:05 +0000 (17:10 +0200)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Sun, 22 Jul 2007 01:37:08 +0000 (18:37 -0700)
It is not fully softirq safe anyways.

Can't do a WARN_ON unfortunately because it could trigger in the
panic case.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
arch/x86_64/kernel/smp.c

index 0694940..df4a828 100644 (file)
@@ -386,9 +386,9 @@ int smp_call_function_single (int cpu, void (*func) (void *info), void *info,
                return 0;
        }
 
-       spin_lock_bh(&call_lock);
+       spin_lock(&call_lock);
        __smp_call_function_single(cpu, func, info, nonatomic, wait);
-       spin_unlock_bh(&call_lock);
+       spin_unlock(&call_lock);
        put_cpu();
        return 0;
 }